NASA Technical Memorandum 105366

THE EFFECT OF ION PLATED SILVER AND SLIDING FRICTION ON TENSILE

STRESS-INDUCED CRACKING IN ALUMINUM OXIDE

Harold E. Sliney and Talivaldis Spalvins National Aeronautics and Space Administration

Lewis Research Center Cleveland, Ohio 44135

ABSTRACT

A Hertzian analysis of the effect of sliding friction on contact stresses in alumina is used to predict the critical load for crack generation. The results for uncoated alumina and alumina coated with ion plated silver are compared. Friction coefficient inputs to the analysis are determined experimentally with a scratch test instrument employing a 0.2 mm radius diamond stylus. Series of scratches are made at constant load increments on eoated and uncoated flat alumina surfaces. Critical loads for cracking are detected by microscopic examination of cross sections of scratches made at various loads and friction coefficients. Acoustic emission and friction trends are also evaluated as experimental techniques for determining critical loads for cracking. Analytical predictions correlate well with the micrographic evidence and with the lowest load at which AE is detected in multiple scratch tests. Friction/load trends are not good indicators of early crack formation. Lubrication with silver films reduced friction and thereby increased the critical load for crack initiation in agreement with analytical predictions.

NOMENCLATURE

elastic modulus, GPa

mean Hertz stress, GPa

radius of hemispherical contact, m

normal static load, N

normal load during sliding, N

Poisson’s ratio

friction coefficient

surface static tensile stress, GPa

surface tensile stress during sliding, GPa

superscript denoting critical value for crack initiation

INTRODUCTION

It has been shown by a number of investigator^'^ that the friction coefficients of unlubricated ceramics are generally high, on the order of 0.5 to 1.0. Wear rates are also higher than might be expected considering the high hardness of many ceramics and the extremely

smooth surfaces to which they can be finished. Ceramics often wear by a microfracture mechanism. The origin of this type of wear is the generation of a crack pattern in the near- surface (surficial) region of the tribological contact. The cracks propagate and intersect at and below the surface to produce loose wear particles. Because ceramics are weak in tension and have very little ductility, the surface cracks are formed by surface tensile stresses. These tensile stresses are a strong function of the tangential friction forces in sliding contacts. Therefore, the high friction coefficients characteristics of ceramics are a primary reason for their high wear during unlubricated sliding. It is to be expected, therefore, that reduction in friction, while desirable in itself, should also alleviate the problem of wear by microfracture.

One way to reduce the friction of ceramics is to coat them with soft, metallic films. Erdemir has shown that ceramics can be coated with adherent, lubricative films of silver by ion beam assisted deposition (IBAI)).~ This is a line-of-sight process that coats only those surfaces that are in the straight line path of the ion beam. The same can be said about other PVD processes such as sputtering although magnetron sputtering has a limited throwing power because of the magnetically curved path of the ion plasma. Ion plating is another PVD process, and it has the advantage of superior “throwing power,” or the ability to deposit films around corners and even on the back of specimens. Direct ion plating cannot be done on electrical insulators such as most ceramic materials. However, as reported by the authors in a companion paper,5 insulators can be coated by placing an accelerating, conductive screen between the filament-heated sources of the coating metal and the ceramic. We call this process screen cage ion plating (SCIP).

In ref. 5 , we showed that SCIP-deposited silver reduced the friction coefficients of alumina. In the present paper, we analytically and experimentally determine the effect of this reduced friction on the critical load for producing cracks and microfracture in alumina. The engineering significance of this is to demonstrate the benefits of reduced friction on the dynamic load carrying capacity of ceramic, sliding contact bearing surfaces.

STRESS ANALYSIS

Model

Microstress studies of static and sliding contacts of ceramics have been treated in the These analyses clearly demonstrate how sliding friction increases surficial stresses

and thereby reduces the critical load, W”, for crack initiation and microfracture damage of brittle materials. Since microfracture is a common wear mode for ceramics, it is of technological interest to quantify the effect of friction coefficient upon the microstresses that are generated in these materials during sliding.

The easiest contact geometry to model for high contact stresses is that of a hemisphere on a flat. A small circular contact area is formed by elastic deformation when a hemisphere is pressed against a flat. For static contacts, the associated stresses are calculated with standard Hertz equations. In our study, the mean static Hertz stress, H,, and the surface tensile stress component, o,, are computed as functions of the normal load on the stylus. Then, the effect of sliding friction on the tensile stress component is calculated. The analyses are discussed in the following section.

2

Static Hertz Stress

The stresses in the concentrated contact of a sphere against a flat are compressive within the contact circle, but are tensile just outside of the circle. The tensile stress is in the radial direction and is a maximum along the circumference of the contact; it decreases rapidly with dis- tance away from the contact. These compressive and tensile stresses are calculated using the elasticity analysis described in Hertzg and in the classic elasticity text by Timoshenko and Goodier." The mean compressive Hertz stress within the static contact of a sphere on flat is given by the equation:

WE2 H, = 0.385

E is the reduced elastic modulus given by:

where E, and E, are the elastic moduli of the materials in contact and v, and v, are the corresponding Poisson's ratios. The values for these properties for diamond and alumina are given in Table I.

The maximum static Hertz stress is 1.5 larger than the mean. The mean is used here for comparison with Vickers indentation hardnesses, which are mean values, and are an upper limit for the valid application of Hertzian elastic analyses.

The scratch test is of course intentionally designed to irreversibly damage (scratch) the surface of the specimen. At the lightest loads, the stylus elastically (reversibly) deforms the surface. As the load is increased, brittle materials such as ceramics crack and fracture. We define the mean Hertz stress at which the first cracks form as the critical mean Hertz stress, HZ. The associated critical tensile stress component and load are 0: and W:.

The mean Hertz stress, H,, for a 0.2 mm radius diamond stylus on alumina was calculated from eqs. (1) and (2) and the elastic properties listed in Table I. Figure 2 gives H, as a €unction of normal loads up to 50 N for the static contact of the diamond stylus on alumina. Corresponding curves for 4.76 mm radius diamond and alumina spherical contacts are also given. This larger radius is commonly used in pin on disk friction and wear tests and is included for reference. The calculated H, is about 18.5 GPa at a stylus load of 50 N. This is above the indentation pressure of 15.7 to 17.7 GPa for alumina derived from published Vickers hardness numbers of 1600 to 1800 kg/mm. 2J4315 We can expect the critical stress for microcrack initiation to be somewhat lower than the indentation pressure. For example, Ikawa and Shimada reported critical stresses for cracking of sapphire in static contact with a 0.2 mm radius diamond stylus to be about 25 percent lower than the indention hardness.6 However, in the following discussion, we will take H: = 17 GPa for the formation of significant surficial damage. The corresponding W: computed from eq. (1) is 38 N.

3

Dynamic Surface Stresses During Sliding

During sliding, the tangential friction force greatly increases the tensile stress component u As previously stated, it is this tensile stress component that is primarily responsible for cracking in brittle materials. Therefore, the dynamic critical load, W:, to initiate cracking is much lower than the static critical load, W:. The following discussion considers the effect of sliding friction on the dynamic load capacity of polycrystalline alumina.

P'

According to Hertzian theory, the maximum static tensile stress, u,, as a function of H, and Poisson's ratio is expressed by the following equation:

1 2

U, = -(1 - 2v)H,

For alumina with a Poisson's ratio of 0.23, U, is 0.27 H, and for the assumed H:, of 17 GPa, u, is 4.6 GPa. This will be taken as U: for alumina throughout this paper. We will next consider the effect of friction on the magnitude of the surface tensile stress during sliding, u P'

Enomoto and Tabor (1981)11 have reported the effect of friction on increasing the tensile stress component during sliding. They quote Hamiliton and Goodman (1966) l2 who showed that the surface tensile stress is increased during sliding by a multiplying factor:

(3)

where, p is the friction coefficient and:

3n (4 + v) 8(1 - 2v)

A =

For alumina, A = 9.2. Therefore,

ap(sliding) = (1 + 9.2p)uo(static)

A plot of the multiplying factor as a function of p is given in Fig. 3. To avoid crack formation, u

P numerical value into eq. (6) must not exceed the crack critical value of 4.6 GPa. When we substitute this

4.6 (1 + 9 . 2 ~ )

6, =

( 5 )

(7)

Therefore, U, and the H, prior to sliding must be less than their critical values in the static case by a factor of 1/(1 + 9 . 2 ~ ) to avoid cracking yhen frictional sliding commences. Because the load is proportional to the third power of H, W

(1/1 + 9 . 2 ~ ) ~ relative to the critical static load, W,. Plots of these two factors as a function of

must be reduced by a factor of P *

4

friction coefficient are given in Fig. 4. They give the fractional reduction in load and associated initial Hertz stress necessary to maintain the same surface tensile stress as in a static contact, and are expressed as the normalized Hertz stress, H /Ho, and the normalized load W /Wo.

The average friction coefficient during constant load scratch tests on alumina is 0.18 at

CL P

loads up tp 20 N. For this case, H /H, and W /W, are 0.37 and 0.051 respectively. In c$her words, W for a moderate sliding &tion coeftcient of 0.18 is only about 5 percent of W,.

c1

For silver plated alumina with a typical friction coefficient of 0.10, W * = 14 percent of P W,. Therefore, ion plated silver should almost triple the allowable load on alumina in sliding

contact with diamond. Computed critical loads and stresses for diamond on uncoated and silver coated alumina are summarized in Table 111.

For higher friction coefficients of 0.5 to 1.0, which are characteristic of most unlubricated ceramics, the detrimental effect of friction would be more*severe than it is for diamond on alumina. For example, at apiction coefficient of 0.6, W = 0.36 percent of WE, and at a friction coefficient of 1.0, W = 0.10 percent of Wz. Therefore, even very lightly loaded

c1 contacts on unlubricated ceramics can be.expected to produce cracks during sliding. The conclusion is consistent with the microfracture wear mode that is characteristic of unlubricated ceramics, especially in concentrated contacts and on rough surfaces where the stresses on asperities are higher than stresses computed for ideally smooth surfaces.

P

EXPERIMENTAL SECTION

Material

The alumina is a sintered, polycrystalline grade of alpha-alumina. Density is 3.78 g/cm 3 or

about 95 percent of theoretical. The specimens are flat blocks that are surface ground to a finish of 0.25 to 0.38 pm. We measured an average Vickers hardness of 1607 kg/mm2. Sasaki14 reports a value of 1800 kg/mm2. This or somewhat greater variability can be expected from various sources of alumina.

The silver coatings are 1 and 4 pm thick.

The stylus used in the scratch testing has a tip radius of 0.2 mm and a cone angle of 120".

Procedure

A schematic diagram of the screen cage ion plating process (SCIP) is shown in Fig. 1. The process is described in some detail in ref. 5. Briefly, the specimen is placed in a vacuum chamber where it is surrounded with a 20 mesh silver screen that functions as an electronic grid. A negative potential is applied to the screen. The evaporation source is the anode. Positively charged ions of the evaporated metal, in this case silver, are accelerated to the screen. Most of them pass through the openings in the screen and are deposited in the neutrally charged substrate within the cage. Since the negative field surrounds the substrate, all surfaces become coated. The surface 180° from the silver evaporation source is coated with about one-fourth the coating thickness of the surface facing the source.

A series of scratches are made with a commercial (Revetest) scratch test machine on uncoated and on SCIP silver-coated alumina specimens. Each scrabch is 10 mm long and is

5

4

made at a velocity of 10 mm/min. The scratches are made with a constant load on the stylus. The lightest constant load is 1.6 N. Loads are then increased in approximately 5 N increments to 50 N. The friction force and acoustic emission are continuously measured as the scratches are made. The specimens are then sectioned in the direction transverse to the scratch directjon, mounted, and polished for metallographic examination to determine the critical load, W surficial damage.

for I.1'

RESULTS

Photomicroscopy

Alumina.-The microstress calculations predict that alumina should undergo microfracture damage at all but the lighter stylus loads. This is verified by the cross section photomicrographs of the scratches on both uncoated and coated alumina (Figs. 5(a) to (c) and 6) which show the effect of stylus load on surficial damage. The observed damage is described qualitatively in Table 11, and is of two types: (1) microfracture of grain boundaries within the Hertzian zone; and (2) radial cracks propagating beyond the Hertzian zone. Both types are illustrated by SEM micrographs in Figs. 7 and 8. The predicted values for the onset of microfracture damage are quantitatively compared to micrographic evidence in Table 111.

Photomicrographs (Fig. 5(a)) of scratch cross sections on alumina show microfracture damage in scratch$s made at 5.5 N, but not at 1.6 N, in good agreement with analytical predictions of W

Silver on Alumina.-Silver coatings of 1 and 4 pm thickness were scratched. Some silver remained in scratches made at the highest loads, but the coatings did not reduce surficial damage at stylus loads above 30 N. However, surficial damage was mitigated by the silver coatings at loads below about 30 N.

= 1.9 N. P

A typical friction coefficient for the 1 pm coating is 0.08. The predicted W * for a friction coefficient of 0.08 is 7.6 N (Hm* = 9.9 GPa). Figure 5(b) shows no damage at 5.5 N, the onset of microfracture damage at 10.7 N, and progressively greater damage at higher loads. The observed W" between 5.5 and 10.7 N agrees with the predicted value of 7.6 N.

c1

The friction coefficient for the 4 pm coating is typically 0.10. The predicted W * is 5.3 N, I.1 but no damage is noticed at 10.7 N, and the onset of damage is seen at 14.7 N. Radial crack

propagation is first observed at 30.6 N (Fig. 6). The increased protection by the thicker coating despite a slightly higher friction coefficient is probably due to the greater load sharing capability of the thicker film. The5efore our calculations, that only take frictional effects into account, accurately predicted W for uncoated alumina and alumina coated with 1 pm of silver, but the correlation is not as good for the 4 pm coating. P

Friction and Acoustic Emission (AEj

Acoustic emission traces and friction traces for uncoated and SCIP silver coated alumina during progressive load scratch tests are shown in Figs. 9(a) and (b). For uncoated alumina, small increases in AE are seen at about 3 N, and a large increase in AE begins at about 6 N. AE for silver coated alumina begins at 9 N and increases with load. The onsets of AE are in fair agreement with the predicted onset of microfracture and microscopy observations. However, the agreement is not as good in all cases. Table I11 gives the scatter in AE results in repeat

6

scratch tests. * The highest loads for the onset of acoustic emission were much higher than the predicted W values. The best, agreement is with the lowest observed AE signals. This is * consistent with our observations during constant load scratch tests at loads greater than W

P that AE pulses do not always occur at the beginning of a scratch nor continuously during the scratch. There is apparently a statistical distribution in the critical loads at which detectable AE signals occur. The predicted values of W

P AE is detected in repeated single pass scratches of the same specimen. are in agreement with the lowest loads at which

The friction traces on Fig. 9(b) show that friction coefficients, (the slopes of the F, versus N traces) are the least reliable indicators for the onset of microfracture damage because friction coefficients do not increase until loads well beyond the onset of microfracture are reached. The increases in slope correlate fairly well with the loads at which radial crack propagation becomes significant, but it is uncertain whether this is a cause and effect relationship or a coincidence.

Comparison of Methods

* Micrographic cross sections of the scratches gives the most reliable evidence of W and the

P best correlation with stress analysis predictions. The disadvantage is the time and expense involved in the preparation and examination of the specimen mounts.

AE in progressive load ttsts is a convenient, quick method. It appears to be a reliable technique for determining W

P statistical distribution of W as a function of load. These lower end values correlate

P reasonably well with prediction and metallographic evidence.

if enough tests are done to establish the lower end of the *

Friction coefficients are the least reliable of the methods used to determipe W * In some P'

P*

cases, the roughness of the friction force (Ft) traces increase at loads near W but* the slopes (friction coefficient, p ) do not show gn upward trend until loads are well above W by Spalvins and Sliney in ref. 5., the friction coefficient of silver coated alumina decreases with increasing load, passes through a minimum, then increases as damage to*the alumina substrate progresses. Because of this behavior, we were not able to determine W force or friction coefficient as a function of stylus load.

P' As shown

from traces of friction P

SUMMARY AND CONCLUSIONS

Silver films were applied to aluminum oxide by the screen cage ion plating (SCIP) method. Scratches were made with a diamond stylus at normal loads from 1.6 to 50 N on the uncoated and silver coated surfaces to experimentally determine the critical loads for the onset of crack formation in the alumina substrate. Experimental results were compared with the predictions of surface stress analyses. The major conclusions from this study are:

1. SCIP coated silver reduces the friction of diamond on alumina and thereby increases * W , the critical load for the onset of microfracture in alumina.

P

2. The validity of the analytical model to predict the effect of sliding friction on W * was P verified by micrograph? of the scratch cross sections made at loads below and above the

predicted values of W P -

7

3. Acoustic emissions can be used to detect cracking in'the alumina as the load is increased. However, in repeated tests, there is considerable data scatter in the load at which the first AE signal is detected. The lowest values in the AE scatter distribution agree with micrographic evidence and analyses.

4. Friction measurements are not useful for detecting early crack generation. Friction coefficients do not generally increase until considerable substrate breakup has occurred.

8

REFERENCES

1. Sutor, P., “Tribology of Silicon Nitride and Silicon on Nitride-Steel Sliding Pairs,” Ceram. Eng. Sci. Proc., 5, pp. 460-469 (1984).

2. Sikra, J.S., Krysiak, J.E., Eklund, P.R., and Ruh, R., “Friction and Wear Characteristics of Selected Ceramics,” Am. Ceram. SOC. Bull., 53, pp. 581-582 (1974).

3. Breznik, J., Breval, E., and MacMillan, N.H., “Sliding Friction and Wear of Structural Ceramics,” J. Mater. Sci., 20, pp. 4657-4680 (1985).

4. Erdemir, A. et. al., “Solid Lubrication of Ceramic Surfaces by IBAD-Silver Coatings for Heat Engine Applications,” Tribol. Trans., 33, pp. 511-518, (1990).

5. Spalvins, T., Sliney, H.E. and Deadmore, D., “Screen Cage Ion Plating (SCIP) and Scratch Testing of Polycrystalline Aluminum Oxide,” NASA TM, 1992. (to be published)

6. Ikawa, N. and Shimada, S., “Microstrength Measurement of Brittle Materials, ” Technology Reports of Osaka University, 31, (1622), pp. 315-323 (1982).

7. Richardson, D.W., Lindberg, J.L., Carruthers, W.D., and Dahn, J., “Contact Stress Effects in Si,N, and Sic Interfaces,” Ceram. Eng. Sci. Proc., 2, pp. 578-588 (1981).

8. Swain, M.V., “Microfracture about Scratches in Brittle Solids.” h o c . Roy. SOC. London, -9 Ser A 366, pp. 575-597 (1979).

9. Hertz, H., “On the Contact of Elastic Solids,” J. Math. (Crelle’s Jour.), 92. pp. 156-171 (1881).

10. Timoshenko, S. and Goodier, J.N., Theory of Elasticity, 2nd edition, McGraw-Hill, (1951).

11. Enomoto, Y. and Tabor, D. “The Frictional Anisotropy of Diamond,” Proc. Roy. SOC. London, Ser. A, 373, pp. 405-417 (1981).

12. Hamilton, G.M. and Goodman, L.E., “The Stress Field Created by a Circular Sliding Contact,” J. Appl. Mech., 33, pp. 371-376 (1966).

13. Chrenko, R.M. and Strong, H.M., “Physical Properties of Diamond,” General Electric R&D Rept. GIDEP-E062-0548 (1975). (Avail. NTIS, AD-B035657L).

14. Sasaki, S., “The Effects of Surrounding Atmosphere on the Friction and Wear of Alumina, Zirconia, Silicon Carbide, and Silicon Nitride,” Wear of Materials, Vol. 1, ASME, pp. 409- 417 (1989).

15. Sliney, H.E. and Deadmore, D.L., “Friction and Wear of Oxide Ceramic Sliding Against IN-718. Nickel Base Alloy at 25 to 800 “C in Atmospheric Air,” NASA TM-102291 (1989).

9

Diamond

Alumina

TABLE 11.-QUALITATIVE OBSERVATIONS OF SCRATCH CROSS SECTIONS BY

OPTICAL MICROSCOPY AT X200

Elastic modulus Poisson's ratio Vicker's hardness

GPa PSI/106 ------ GPa kg/mm

"1020 148 b0.20 "88 8333

'363 53 '0.23 18 '1800 d1600

2

[Legend: C: Propagated cracks, D: Grain boundary damage.] I

Uncoated alumina Nominal load, N

1 pm Ag/alumina

- ~~

1.6 5.0

10 15 20 25 30 35 40 50

No damage Slight D Slight D

Moderate D One C, mod. D Moderate C, D

Increased C & D Severe C & D Severe C & D Severe C & D

No damage No damage

Slight D Slight D

Moderate D Moderate D

One C, mod. D Increased C & D

Severe C & D Severe C & D

4 pm Ag/alumina

No damage No damage No damage

Slight D Moderate D Moderate D

One C, mod. D Increased C & D

Severe C & D Severe C & D

10

TABLE 111.-CALCULATED AND OBSERVED CRITICAL STYLUS LOADS FOR 0.2 ~ll~~l

RADIUS DIAMOND ON ALUMINA

Silver coating

thickness, Pm

None (Static)

None 1 4

Friction coefficient,

P

0

.18

.08

.10

Calculated

Critical mean Hertz stress

GPa €I* 7

17

6.3 9.9 8.8

Critical load W * ,

N

38

1.9 7.6 5.3

Critical Load W*

Experiment a1

Micrographic, N

1.6 > 5.5 5.5 > 10.7 10.7 > 14.7

Acoustic emission

Number of tests

Scatter, N

36-52

2-9 6- 10 5-17

,J screen-cage 1 (cathode)

\,Ceramic substrate

Resistance-heated tungsten boat containing silver evaporant (anode)

Filament power ---I--- supply

Average , N

42

4.5 7.8 9.3

Figure 1 .-Schematic for screen-cage ion plating of ceramics.

11

20 - ,-- Indentation pressure, /// Pm, of AI203

0.2 mm, diamond -, 2 15- CII

E I vi v) W' 2 10 -

t I

1

2 .If 4.76 mm, diamond 7,

2 -

I I I I I l l l l I I I 1 1 1 1 1 1 I I I I 1 1 1 1 1

II 4.76 mm, A 1 2 0 3 7 , '\, . . I I I

0 10 20 30 Normal load, W, Newton

(a) Linear scale.

100

40 m

0 .2 .4 .6 .8 1.0 Friction Coeff., I.I

Figure 3.-Effect of dynamic friction coefficient on the tensile stress component in a ball on flat hertzian contact on alumimum oxide.

u 40 50

laO .8 A 20

f vi

10 2 " 8 v)

6

I

5 U

.6

k

.2

0 .2 .4 .6 .8 1.0 Friction coeff., p

Figure 4.-The effect of friction coefficient on reducing the critical load, W*, and initial hertz stress, Ha, to produce tensile cracks during sliding of a hemisphere on an alumina flat.

12

ORlGlNAL PAGE BLACK AND WHITE PHOTOGRAPh

1.6 N

10.5 N

1.6 N 1.6 N

5.5 N

10.7 N

15.4 N

20.2 N

(a) Uncoated alumina

, - _ - .. +. - , . 20.7 N

(b) 1 pm Silver on alumina

4.7 N

10.7 N

14.7 N

20.7 N - 50 pm

(c) 4 pm Silver on alumina . .

Figure 5.4rogressive microfracture with increasing load.

13

30.6 N U 50 pm

30.6 N

. _

50 N 50 N

Uncoated alumina 4 pm Silver on alumina Figure 6.-Progressive microfracture and radial crack propagation at high loads.

Figure 7 . C E M micrograph of microfracture damage in the hertzian zone under a scratch made on alumina at a 5.5 N stylus load.

(a) Microfracture and radial crack damage.

(b) Radial crack initiation and propagation site.

Figure 8 . 4 E M micrographs of severe damage under a scratch made on alumina at a 30.5 N stylus load. ,

(a) Acoustic emission. 0

30r P 20 -

z w- U

Y

0 20 40 60 80 Normal load, W, Newtons

(b) Friction.

Figure 9.-Effect of "SCIF" silver coating in scratch tests of alumina.

15

REPORT DOCUMENTATION PAGE

___

4. TITLE AND SUBTITLE

The Effect of Ion Plated Silver and Sliding Friction on Tensile Stress-Induced Cracking in Aluminum Oxide

6. AUTHOR@)

Harold E. Sliney and Talivaldis Spalvins

7. PERFORMING ORG@ZATION NAME(S) AND ADDRESS(ES)

\National Aeronautics and Space Administration : :r iLFwis Research C&$er:, (i Cleveland, Ohio 441i3sT3191

. I .

Form Approved OMB NO. 0704-0188

9. SPONSORINGIM~~IZ~RING-AGENC~ NAMES(S) AND ADDRESS(ES) 1 - I

1. AGENCY USE ONLY (Leave blank)

1

I

National Aeronautics and Space Administration Washington, D.C. 20546 - 000 1

I .

2. REPORT DATE 3. REPORT TYPE AND DATES COVERED

1991 Technical Memorandum 5. FUNDING NUMBERS

14. SUBJECT TERMS

Ceramics; Surface stresses; Indentation stress; Friction stress; Ion plating of ceramics

WU - 505 -63 - 5A

15. NUMBER OF PAGES 16

16. PRICE CODE

8. PERFORMING ORGANIZATION REPORT NUMBER

17. SECURITY CLASSIFICATION OF REPORT

Unclassified

E- 6651

18. SECURITY CLASSIFICATION OF THIS PAGE

Unclassified

IO. SPONSORING/MONITORING AGENCY REPORT NUMBER

NASA TM - 105366

11. SUPPLEMENTARY,NOTES. :: ; Prepared for the Annual Meeting of the Society of Tribologists and Lubrication Engineers, Philadelphia, Pennsylvania, May 4-7, 1992, Responsible person, Harold E. Sliney, (216) 433-6055.

12a. DISTRIBUTION/AVAILABlLlTY STATEMENT I12b. DISTRIBUTION CODE t

Unclassified -Unlimited Suljject Category 27

I 13. ABSTRACT (Maximum200 words)

A Hertzian analysis of the effect of sliding friction on contact stresses in alumina is used to predict the critical load for crack generation. Tlie results for uncoated alumina and alumina coated with ion plated silver are compared. Friction coefficient inputs to the analysis are determined experimentally with a scratch test instrument employing a 0.2 mm radius diamond stylus. Series of scratches are made at constant load increments on coated and uncoated flat alumina surfaces. Critical loads for cracking are detected by microscopic examination of cross sections of scratches made at various loads and friction coefficients. Acoustic emission and friction trends are also evaluated as experi- mental techniques for determining critical loads for cracking. Analytical predictions correlate well with the micro- graphic evidence and with the lowest load at which AE is detected in multiple scratch tests. Frictionfload trends are not good indicators of early crack formation. Lubrication with silver films reduced friction and thereby increased the critical load for crack initiation in agreement with analytical predictions.

OF ABSTRACT Unclassified

Standard Form 298 (Rev. 2-89) Prescribed by ANSI Std. 239-18 298-102

National Aeronautics and Space Administration

Lewis Research Center Cleveland, Ohio 44135

Official Business Penalty for Private Use $300

FOURTH CLASS MAIL

ADDRESS CORRECTION REQUESTED

Postage and Fees Paid National Aeronautics and Space Admlnistralion NASA 451